Image Cache

Stephen Lime steve.lime at dnr.state.mn.us
Thu Nov 4 14:22:21 EST 1999


You could hack mapserv.c and have it spit out
the tag you mention. Another way would be to
come up with a more unique naming convention.
Version 3.3 now uses time + pid for what should
be a truly unique name. If you can hold out for a
few more days your problem should be solved,
assuming it'll compile on NT. ;-)

Steve 

<<< "KIERAN AMES" <kames at keyspanenergy.com> 11/ 4  1:04p >>>
I've installed a port of Mapserv onto a NT4.0 machine and am now
experimenting. I've run across an implementation problem with regard to
images being cached.

I believe that image names are created with the map name concatenated to
the processID. In NT, process IDs don't run very high and are re-used.
Therefore, when a new image is created, the client browser sees an image
name and if it finds it in its local cache, it doesn't fetch the revised
image that was just created.

I can control this in regular cgi (Perl) by sending non-parsed headers
and delivering a "Pragma:No-cache" line in the http headers. Is there an
equivalent solution with MapServ on an NT platform?

Thanks very much.
Kieran Ames

PS. I'm new to MapServ, but I think its GREAT!






More information about the mapserver-users mailing list